Understanding the microvascular changes of diabetes is crucial to the development of improved therapy. In the diabetic retina, areas of ischemic tissue are thought to lead to deficient oxygenation and the production of vascular endothelial growth factor, which enhances vascular permeability and plays a major role in retinopathy. Our preliminary data in an animal model of diabetes (streptozotocin injection) demonstrate early arteriolar constriction in the initial weeks of hyperglycemia. Experiments from our lab suggest a localized mechanism of microvascular dysfunction, in which mediators derived from inflammatory cells diffuse from venules to closely paired arterioles to induce vasoconstriction, either directly or through an attenuation of the vasodilator nitric oxide. In a related model (microvascular dysfunction in the mesentery of diabetic rats), we previously have demonstrated a substantial attenuation of nitric oxide in arterioles closely paired with postcapillary venules. We hypothesize that the same could be true in the diabetic retina, and that the attenuation in nitric oxide is related to the increase in reactive oxygen species such as superoxide. Moreover, we hypothesize that thromboxane derived locally from inflammatory cells contributes significantly to the arteriolar vasoconstriction. We have obtained exciting preliminary data (in the retina of both mice and rats) indicating that inhibition of thromboxane synthase reverses the arteriolar vasoconstriction induced by diabetes. Thromboxane is a highly potent vasoactive molecule, and can induce vasoconstriction directly by binding to its receptor on vascular smooth muscle cells. In addition, the vasoconstrictor has been found to contribute to oxidative stress, and inhibition of thromboxane acutely elevates nitric oxide bioavailability. Models of streptozotocin-induced diabetes demonstrate retinal hypoxia, increased cell death, an increase in VEGF production, and an increase in vascular permeability. We propose that vasoconstriction contributes to these deleterious consequences, and that improvements in these endpoints can be accomplished via inhibition of thromboxane and reactive oxygen species. Our specific aims are to investigate the role for thromboxane and reactive oxygen species in the early retinal arteriolar constriction induced by diabetes, and to determine whether inhibition of vasoconstriction improves endpoints of diabetic retinal complications.